Conversion optimisation with Behavioural Design: why A/B tests aren’t enough

Picture this: your e-commerce team has been testing for three months. Button colours, headlines, image positions, the number of fields in your form. The conversion rate has climbed from 4.1% to 4.8%. Everyone is pleased. But nobody asks the question that actually matters: what was that 95.2% doing at the precise moment they left?

That is the blind spot of classic conversion optimisation. You are testing treatments. But you have never made a diagnosis. And without a diagnosis you are managing symptoms, not solving the real problem.

A conversion rate of 5% means 95% of your visitors said no. The question is: why?

Classic CRO is symptom management

Compare it to a doctor who prescribes a painkiller to every patient without any examination, because painkillers work on average. Sometimes the patient has a broken leg, sometimes an infection, sometimes psychosomatic symptoms. The painkiller dampens the signal. The problem stays.

That is how classic conversion optimisation works. You have a page with low conversion. You test variant A against variant B. Variant B wins by 0.7 percentage points. You roll out variant B and start the next test. But you still have no idea why that 95% dropped off. Was it fear of commitment? Was the benefit of your product not made concrete enough? Was the status quo simply comfortable enough? Or was the form too long?

That question is not trivial. Because if it is fear of commitment, a shorter headline does nothing. You need a risk guarantee. If it is choice stress from too many options, a better call-to-action colour does nothing. You need progressive disclosure, or a radical simplification of your offering. The intervention depends entirely on the diagnosis.

You cannot design the right intervention for a problem you have not diagnosed. A/B testing gives you results. Behavioural Design gives you understanding.

The conversion barrier is almost always psychological

This is the insight most CRO teams miss. They look for the cause of low conversion in the interface: too many fields, the wrong order, slow load times. These are real factors. But the dominant reason people do not convert is psychological.

Dan Ariely described in Predictably Irrational (2008) how people respond to choice situations in consistently irrational ways.^[1] We have no good intuitions for absolute value. We compare, relatively, against a reference point. We are paralysed by too many options. We procrastinate when the gain is too abstract. We avoid any step that suggests loss of freedom. These are not bugs in human behaviour. They are the rules.

Once you know those rules, you look at a low conversion rate differently. You no longer see a technical problem. You see a behavioural problem with a psychological cause. And that is where the interesting work begins.

Three scenarios: the diagnosis A/B tests never make

The e-commerce team that had been testing for three months

Take a team responsible for the product page of a mid-sized brand with a reasonable assortment. They have been testing for months. Headlines, image sequences, the length of product descriptions. Results are disappointingly flat. Then they decide to observe users instead of testing them. They watch session recordings. And they see it.

Visitors scroll through fifteen product options on a page, click one, read the description, go back, click another, scroll again. They decide nothing. They leave.

This is not a copy problem. This is choice stress. Barry Schwartz described in The Paradox of Choice (2004) how more options lead to less satisfaction and fewer decisions.^[2] The anxiety here is not fear of a bad purchase. It is the cognitive load of comparing. The comfort is postponement: I will decide later.

The intervention the team then designs has nothing to do with button colours. They introduce progressive disclosure: a filtering mechanism that guides visitors with two questions to the three most relevant options. And they add specific social proof: not “900+ reviews” but “Chosen by 2,847 professionals who also searched for X.” The conversion rate rises 23% in six weeks. Not through testing. Through diagnosing.

The HR team that communicated better and got fewer sign-ups

An HR team launches an internal development programme. They invest in a professional communications campaign. Good visuals, a clear email, an intranet page. Sign-ups fall well short of expectation. They revise the campaign. Another email. A reminder. Nothing moves significantly.

The mistake they are making is the same one most campaigns make: they communicate the content of the programme, not the gain for the participant. And they expect the recipient to take the step of signing up unprompted. But people do not sign up for programmes simply because those programmes are good. They sign up when the gain is concrete enough to move them, and the barrier is low enough to overcome.

Two things change the situation. First: an implementation intention prompt in the sign-up email. Not “Sign up if you would like to participate”, but: “Pick a date now. Put it in your calendar. Sign up.” Research by Peter Gollwitzer showed as early as 1999 that specifying when, where and how you will do something more than doubles the likelihood that you actually do it.^[3]

Second: the default changes. Instead of opt-in, the organisation sends a message telling employees they are already registered for an orientation session, with the option to opt out. The result: participation rises from 14% to 61%. Not because the programme improved, but because the choice architecture changed. Richard Thaler and Cass Sunstein described this mechanism in Nudge (2008): the default option disproportionately determines the outcome.^[4] You can read more about how defaults work as a choice intervention in that article.

The SaaS company that gave away free trials but made no paying customers

A B2B SaaS company has a free trial that performs well on starts: 40% of visitors who see the page begin a trial. But only 6% of trial starters convert to a paid subscription. The team A/B-tests the upgrade page. Nothing moves significantly.

They do something most SaaS teams never do: they call ten people who started the trial but did not convert. No survey. A phone call. And they hear two things they did not expect.

First: the gain of the product has not become concrete enough during the trial. People used the tool superficially, never reached the moments of insight that make it valuable, and ended the trial without being able to say what it had given them. The gain had remained too abstract.

Second: upgrading felt like an irreversible commitment. Even at an annual price of one hundred pounds per month, the subscription felt like a definitive choice that required certainty they did not have.

The interventions they design target precisely these two forces. They redesign the trial onboarding as a structured journey to the first concrete result, with a checklist that leads users through the three actions that make the product’s value most visible. This is peak-end design: you design the memory of the trial, not just the trial itself.

And they add a risk-reversal guarantee to the upgrade page: the first two months fully refundable, no questions asked. The anxiety about commitment dissolves. Conversion from trial to paid rises from 6% to 19%. Three times higher. With the same button colour as always.

The SUE | Influence Framework as a diagnostic model for conversion

What all three scenarios share: they began with a diagnosis of the four forces that drive behaviour, rather than launching another test. This is precisely what the SUE | Influence Framework does. It maps which psychological forces keep a person in their current behaviour (the visitor who does not convert) and which forces can move them towards the desired behaviour (the purchase, the sign-up, the upgrade).

Take a typical e-commerce product page. Why does that visitor not convert? On the side of the driving forces: the pain of the problem your product solves may not be framed clearly enough. Or the gain, the concrete improvement the buyer experiences, is too abstract, too generic, too distant in the future to activate the System 1 brain.

On the side of the restraining forces: the comfort of the status quo. The visitor currently has something that also solves the problem, perhaps less well, but well enough. That is a powerful anchor. And then there are the anxieties: fear of overpaying, fear that the product does not do what it promises, fear of making an irreversible decision with insufficient information. Each of those forces calls for a different intervention.

If the pain is not being felt sufficiently: frame the cost of inaction. How much is your visitor losing per month by not solving the problem? What does the status quo actually cost them?

If the gain is too abstract: make it concrete. Not “improve your productivity”, but “save 4.5 hours per week on reporting.” The ING research showed that one concretisation question in an enrolment form raised sign-ups by 20%. The same principle applies to conversion: make the future gain vivid and specific.

If comfort is the barrier: make switching easier than staying. Not through persuasion, but by removing friction. This is how choice architecture works as an intervention tool.

If anxiety is the barrier: resolve it directly. A money-back guarantee addresses the fear of a bad purchase. Specific testimonials from people who resemble the buyer resolve the fear that “this won’t work for someone like me.” And a simple, step-by-step form resolves the cognitive anxiety about a complicated process. Airbnb did exactly this with its first major intervention: hiring professional photographers for listings. Not to change the apartments, but to remove the anxiety that reality would differ from the photo.

You change the decision environment, not the person

Here Behavioural Design touches something that classic CRO overlooks entirely. Almost all conventional conversion optimisation tries to persuade the person: better arguments, better copy, better images. But people are not text-processing machines. They respond to context. They respond to defaults. They respond to what others are doing. They respond to how a choice is framed.

This is Kurt Lewin’s insight, which laid the foundation for the Influence Framework: behaviour does not change by making the person different, but by changing the situation. In conversion optimisation that means: design the decision environment so that the desired behaviour is the path of least resistance.

Amazon removed the “Register” button and replaced it with “Continue.” No new arguments. No new images. Just the removal of a barrier at the exact moment of decision. Result: 45% more revenue, worth 300 million dollars in the first year. That is not an A/B test producing a variant that wins by 0.7 percentage points. That is a behavioural diagnosis that identifies and removes the real barrier.

Persuading people to convert works less well than removing what stops them. That difference is the difference between karate and judo.

The same principle applies to every conversion point. With forms: remove fields. Every extra step is a point where behaviour can break off. With checkouts: save card details. Make return trivially easy. With subscriptions: make the free version good enough to demonstrate value, and make the upgrade the logical next step, not a sales conversation.

From testing to diagnosing: how to do it

The practical step is less complicated than it sounds. You do not start with a test. You start with questions to people who did not show the behaviour.

Talk to ten people who visited your page but did not convert. Use behavioural interviews, not surveys. Ask about the past: what exactly did they do at the moment they left? What were they thinking? What made them uncertain? What would they have wanted to know? Intention questions give you rationalisations. Behaviour questions give you reality.

Then map the four forces. Which pains were there that were not framed strongly enough? Which gains were too abstract? Which comforts anchored the visitor in the status quo? Which anxieties blocked the step?

Only once you have that diagnosis do you design an intervention. And the intervention is always aimed at one specific force. Not a generally better page. A specific answer to a specific blockage.

This is the core of Behavioural Design as an approach to conversion: no more guessing, no more hoping for a statistically significant result after a thousand sessions. But understanding what is happening in your visitor’s mind at the moment of decision, and redesigning the environment so that the desired behaviour becomes easier than the alternative.

Frequently asked questions

What is the difference between classic CRO and Behavioural Design?

Classic CRO tests symptoms: which button colour, which headline, which order of elements. Behavioural Design first establishes a diagnosis: why is the visitor saying no? That diagnosis runs through the SUE | Influence Framework, which maps the four psychological forces (pains, gains, comforts and anxieties) that form a conversion barrier. Only once you know which force is blocking behaviour do you know which intervention to design.

Why does a good landing page still convert poorly?

Because most conversion barriers are psychological, not technical. A page can load flawlessly, be perfectly worded and have a clear call to action, while the visitor still drops off. The reason is not in the page itself, but in the fear of commitment, the lack of concrete gain, the presence of better-seeming alternatives in the visitor’s perception, or simply the friction of the decision moment. These are behavioural barriers that a different button colour will not solve.

What is the PGCA analysis in conversion optimisation?

PGCA stands for Pains, Gains, Comforts and Anxieties: the four forces from the SUE | Influence Framework. In conversion optimisation you map which of these four forces forms the barrier at each conversion point. Is it a pain that is not being felt strongly enough? Is there insufficient gain? Is there a comfort anchoring the status quo? Or do anxieties block the step? Only then do you know which intervention raises conversion.

How do defaults work in conversion optimisation?

Defaults are one of the most powerful tools in conversion optimisation. Richard Thaler and Cass Sunstein showed that the default option in a choice situation has a disproportionately large effect on behaviour: people choose what is already set for them. In the case of organ donation, the shift from opt-in to opt-out led to an increase from around 20% to nearly 100% donor consent. You apply this principle directly to forms, subscription options, training sign-ups and checkout flows.

How do you practically apply Behavioural Design for higher conversion?

Start with the diagnosis: interview or observe people who visited your page but did not convert. Ask about their past experiences, not future intentions. Map the PGCA analysis: which force held them back? Then design an intervention that precisely targets that force: anxiety reduction via a guarantee, progressive disclosure to reduce choice stress, specific social proof, or an opt-out default for the desired behaviour.

Conclusion

A 5% conversion rate is not a statistical given. It is a signal that 95% of your visitors were being held back by something at that moment. And that something is almost never a wrong button colour. It is a psychological force: too little pain to move, too abstract a reward, too comfortable a status quo, or an anxiety that was never resolved.

Behavioural Design gives you what A/B testing never can: a diagnosis. And with a diagnosis you can design an intervention that addresses the real barrier. That is the difference between a 0.7 percentage point gain and 23%. Between six months of testing and six weeks of results.

Want to learn how to apply the Influence Framework to conversion and behaviour challenges in your own context? The Behavioural Design Fundamentals training teaches you how to make the diagnosis and design the interventions. Rated 9.7/10 by 10,000+ alumni from 45 countries.

PS

At SUE our working rule is that behaviour that does not change always has a reason. That reason is almost never in the person. It is in the situation, in the choice architecture, in what is missing at the moment of decision. If you take that seriously, you stop testing what works and start understanding why it does not. That is a fundamentally different question. And it produces fundamentally better answers.